| dc.description.abstract | With the increased development of communications, including wire、wireless and network etc., all the multimedia files need compression to be transmitted on a limited bandwidth. In digital audio coding, MP3 is the most popular audio compression format in current days. However, with the progress of coding technology, new standards have been proposed which have higher efficiency and higher quality. MPEG-2 AAC, a new standard proposed by MPEG in 1994, has higher compression efficiency than MP3 and has chance to replace MP3 in the near future. MPEG-4 AAC, a newer standard proposed in 1998, which is based on the techniques used in MPEG-2 AAC but added some more efficient tools. Since the properties of higher compression rate and higher quality, MPEG-2/4 AAC will replace MP3 as the most popular audio compression standard.
Currently, most audio decoder including MP3, AAC are implemented based on the Digital Signal Processor (DSP). However, since the complex algorithms in audio standards, the designers can usually optimize the decoder in C-level. They cannot consider the DSP hardware architecture and algorithms simultaneously. Therefore, the decoding efficiency is decreased and power consumption is increased. In this project, we will implement the MPEG-2/4 AAC audio decoder on PAC DSP with assembly language. First, we will analyze the decoding algorithms and proposed fast algorithms for the initial optimization. Then, we write assembly for further optimization based on the PAC DSP’s architecture and instruction set. Finally, we will complete the optimized design of MPEG-2/4 AAC decoder. | en_US |